Introduction

The data downloaded provides comprehensive match and player information from League of Legends, one of the world’s most popular multi-player online battle arena (MOBA) games.

It features data from 35,000 matches and over 78,000 player records, covering various in-game aspects such as champion choices, performance metrics, match results, and other game play statistics.

The data is organized into7 CSV tables: 1. MatchStatsTbl 2. TeamMatchStatsTbl 3. MatchTbl 4. RankTbl 5. ChampionTbl 6. ItemTbl 7. SummonerMatchTbl.

Data excluded: Summoner data due to privacy restrictions, the owners of the data did not share.

League of Legends Terminology

Champions(Champs) : The characters that players play.
Queue Type: The Game mode. ( 5v5, 2v2 1v1)
Lane: Players dictate which lane they would like to play in certain maps such as classic.
Rank: Level of player, Higher rank better player.
Minions: little soldiers, each team gets little soldiers.
K/D/A: Kills, Death, Assists

Functions used

dplyr : Used for data cleaning, selecting, filtering..
ggplot: Plotting density, bar, jitter plots, smooth curves, faceted graphs
plotly: Interactive graphs
Kable Extra: Scrollable tables and better visualization of tables
Grid Extra: Combining multiple grids
Joined functions: Used to build combined tables

Combined (Joined) Data

What we did is join most of our data sets to have a more efficient data structure. The following data sets were created:

ChampionTbl is the original dataset.

match_rank_tbl: This table displays every match in EU region with the following information: Queue type, Rank, Game duration.

TMR_Tble: Team match rank table is a combined data of one original csv file and a combined data set. TeamMatchTbl and match_rank_tbl. Shows detailed team statistics.

SMR_Tbl: This table displays the players key detailed match stats.

Team_Winner_Tbl: This table only shows teams that have won from TMR_Tble.

MSI_Tbl: This table set will display the players detailed match stats which include kills for both enemy champions and monsters affiliated and include item purchase and summoner spells.

Lets Look at Team Stats!

Why is this interesting?

Well in league of legends, all sorts of kills are important.

Gold is the biggest objective, and it is mainly gained through all sorts of minion kills and enemy champion kills, this allows for solo lane or paired players to buy items that are necessary to get stronger.
Baron kills gives your whole team a chunk of gold and advanced stats for a period of time, leading to a higher chance of winning, dragons on the other hand do not provide gold but do enhance stats.
Tower Kills are important because you get gold and you get an advantage around the map, providing map pressure by allowing your minions to move up without restriction.

Fun fact there are several types of dragons: Infernal, Ocean, Mountain, Chemtec, Hextech, Cloud and Elder dragon.

Mean of Kills, Baron Kills, Rift Herald Kills, Dragon Kills, Tower Kills

Why is it that the mean for Baron, Rift and Dragon Kills is smaller than regular kills?

Dragons spawn every 5 to 7 minutes throughout the game.
Rift Herald first spawn is at 15 minutes.
Baron’s first spawn is at 25 min, if baron is defeated, the boost last about 3 mins. Then afterwards the baron will spawn every 5 to 6 minutes.
Game duration plays a big part in Summoners Rift Map (CLASSIC) and other maps as well such as ARAM.

Fun fact: The first Baron must be defeated with at least 3-4 players, or else you will die. Afterwards 2 players is enough but it takes time to defeat Baron, not an easy tasks to do alone.

Looking at Dragons vs Baron vs Rift Herald, we can see that Dragons are killed more frequently. Spawning is a factor, also you do not need many members to kill a dragon, unless it is the elder dragon.

We can see that Aram and Classic both hold high mean kills, why is that?

Max of Kills, Baron Kills, Rift Herald Kills, Dragon Kills, Tower Kills

Max kills vary through map, as you can see not that many kills in swiftplay, some maps have shorter and more difficult objectives during game, making it more difficult to kill.

We did not include ARAM for Baron, Dragon and Rift Herald kills. - This is simple is because they do not have those monsters in that type of QUEUETYPE - Also you might see CHERRY missing, this is also important to point out, they are an arena type of map. Their teams consist of 2 players instead of the typical 5.

Fun Fact : Jungler usually kills the first dragon with the help of bottom and support.

#Team Game Duration

Lets go ahead and evaluate the Blue teams and Red teams

Blue Team - In Aram, Blue teams seem to require more kills to win a game. - When Blue team has a lead in kills and the game duration is not long, they have a chance to win.

Red Team - The most difficult map for red team is ARAM. If they do not have a lead in kills early on they lose the game. - Same goes for CLASSIC queue types, red team wins when they have more kills early on.

What do we see?

## # A tibble: 7 × 3
## # Groups:   Winning_Team [2]
##   Winning_Team QueueType  Wins
##   <chr>        <chr>     <int>
## 1 Blue         ARAM         13
## 2 Blue         CLASSIC     341
## 3 Blue         SWIFTPLAY     5
## 4 Red          ARAM       4384
## 5 Red          CLASSIC   23804
## 6 Red          SWIFTPLAY   406
## 7 Red          ULTBOOK      11

The following plot shows that most won games are from the red team. The kills are a combined total of both the winning and losing team.

Fun Fact: Some champions are better at early stages of the game. Basically as the game duration continues the champion becomes more and more useless.

Winning Teams

Lets look at the winning teams more deeply.

This table shows that the Blue team has more overall kills than the Red team.

## # A tibble: 2 × 2
##   Side   Kills
##   <chr>  <dbl>
## 1 Blue  579838
## 2 Red   564524

What does this density curve mean? - The height of the line tells you how common a particular game duration is. - Taller areas = more games lasted around that length. - Shorter areas = fewer games lasted around that length

Fun Fact: Summoners Rift (CLASSIC) sometimes last 50 minutes or longer.

Lets look at Player Stats!

Its amazing when you are stacked.

A players stats are important, they can break or make a team.

Inflicting damage is great to look at, if you don’t make enough damage, your hits are basically non existent to an enemy that is stacked.
Killing minions is very important, you get gold by killing them, and gold is amazing.
Dealing damage when you have the right items, basically unstoppable.
Taking damage not so great but can be reduced.

Fun Fact: You basically can die from 3 to 4 consecutive turret hits in the early stages of the game. The damage increases by 40% for each shot (up to a maximum bonus of 120%)

A players role is fundamental in scaling and advancing within the game. - A Classic game consists of 5 different roles (Top, Mid, Jungle, Bottom, Support)

Why is it that Top lane on average deals more damage, kills, and receives more gold?

In a Classic queue there are 3 lanes (Top, Middle, Bottom) but there is also a jungle in between those lanes. Now you only see 4 different paths so where does 5th role go ? Well they go bottom, Bottom and support are always together.

Fun Fact: ARAM maps are only one lane, almost like a free for all. The map forces you to team fight as soon as the game starts.

Player Lane Analysis

Kills vs Lane - Top Lane has the most kills over any other lane - Middle and Bottom differentiate about 2k kills. - Support having the least kills - None is taken into consideration based on other maps but even then it is not much more.

Why is it that NONE doesn’t have the same amount of kills as other lanes? - This is due to queue type. queue goes hand in hand with game duration. The longer the game, the more kills. - Later on we will see how selection of queue type alone affects this kill distribution.

Analyze Gold and Vision Score

GOLD, GOLD, GOLD! The more money you have the better the items. Better items means more damage dealt specially if you are a bottom, Top and Jungle.

Winning teams have a slight higher vision score than losing teams.

How does vision score correlate to winning?

Interestingly enough, vision score has a big impact on kills and deaths.

Vision leads to map control by providing information on enemy positions, which helps the player and team to make better decisions.

Fun Fact: Wards are placed throughout the map to gain vision.

Top 10 Items and Champions

Items are fundamental to scaling a champion. Some items are a non negotiable and are needed, such as boots.

## # A tibble: 10 × 2
##    ItemNames                Total_Items
##    <chr>                          <int>
##  1 Ionian Boots of Lucidity        5886
##  2 Infinity Edge                   5881
##  3 Sorcerer's Shoes                5575
##  4 Berserker's Greaves             5202
##  5 Doran's Blade                   4867
##  6 Mercury's Treads                4774
##  7 Rabadon's Deathcap              4653
##  8 Plated Steelcaps                4146
##  9 Shadowflame                     3764
## 10 The Collector                   3317

## # A tibble: 10 × 2
##    ItemNames             Total_Items
##    <chr>                       <int>
##  1 Anvil Voucher                   1
##  2 Elixir of Avarice               1
##  3 Enhanced Lucky Dice             1
##  4 Minion Dematerializer           1
##  5 Bravery Voucher                 2
##  6 Chemtech Putrifier              2
##  7 Kalista's Black Spear           2
##  8 Lifeline                        2
##  9 Elixir of Force                 4
## 10 Elixir of Wrath                 5

## # A tibble: 15 × 3
##    ItemNames           Items_used_avg Items_used_median
##    <chr>                        <dbl>             <int>
##  1 Abyssal Mask                   260               260
##  2 Aegis of the Legion            193               193
##  3 Aether Wisp                    289               289
##  4 Amplifying Tome               2172              2172
##  5 Anathema's Chains               48                48
##  6 Anvil Voucher                    1                 1
##  7 Archangel's Staff               58                58
##  8 Ardent Censer                  619               619
##  9 Armored Advance               1311              1311
## 10 Atma's Reckoning               134               134
## 11 Axiom Arc                      804               804
## 12 B. F. Sword                   1144              1144
## 13 Bami's Cinder                  101               101
## 14 Bandleglass Mirror             255               255
## 15 Banshee's Veil                 338               338

Champions in the data below show that the most popular are from the bottom lane. These are called AD carry’s, they tend to be always in meta and are popular among different rank levels.

More about the players

It is interesting to see the amount of players that get online to play League of Legends.

A lot of players fall under Master for Rank, almost double the amount of emerald the second top rank.

Classic or Summoners rift is the top most played map. Why do we think that is?

Fun Fact: Summoners rift has two queue styles. Normal and Ranked, Normal is a more casual game play and it does not affect your rank, while ranked is very competitive and tracks players progress by promoting or demoting their rank.

Conclusion

The data extracted from the League of Legends EU region, shows that teams or players who earn more gold, deal more damage, and have higher vision scores generally have a greater chance of winning. More gold allows players to buy stronger items, which improves their performance in fights. Players can choose the same champion as others, so champions are not tied to a single player.

Game duration is strongly linked to kill counts for both Red and Blue teams. As matches last longer, the total number of kills consistently increases, especially for the winning team. Although different queue types vary in their typical match length, the overall pattern remains the same: longer games result in more kills. A single player can participate in multiple queue types, and they can switch between them freely with no limitations.

Overall, the data set highlights the importance of strategic decision-making, teamwork, and individual performance in League of Legends. Success is not determined by a single factor, but rather a combination of gold management, damage output, vision control, champion selection, and effective coordination.

LOL_Dataset_Project

About the dataset